Morphological and Behavioral Evolution in Forest Deer Mice

Abstract

To understand the process of adaptation to novel habitats, it is important to test the genetic and functional mechanisms that may lead to associations both among traits as well as between traits and the environment. Deer mice (Peromyscus maniculatus) consistently vary in multiple traits across habitats, and many of these differences may be due to local adaptation. In particular, forest and prairie ecotypes that differ in pigmentation, tail length, hindfoot length, and climbing behavior have repeatedly evolved in North America. Here, we combined forward genetic approaches in the laboratory with studies of populations in the wild to understand associations among genotype, phenotype, and environment in forest and prairie deer mice. In Chapter One, we tested the extent of phenotypic similarity in two forest-prairie subspecies pairs. We found striking similarities within ecotypes from divergent eastern and western lineages: laboratory-reared forest and prairie mice differed consistently in multiple aspects of skeletal morphology, climbing tendency, and climbing performance in an assay of arboreal locomotion, but not in overall exploratory behavior. A simple analytical model suggested that tail length differences could affect performance by correcting body roll. In Chapter Two, we used a forward genetic approach to determine that most of the characteristic forest traits were not co-inherited in forest-prairie hybrids. We identified several regions of the genome in which forest ancestry is associated with forest-like morphology and determined that most forest phenotypes are likely polygenic. Next, we took advantage of these hybrid mice, in which traits were shuffled, to test for an association between morphology, behavior, and performance. We found that lighter mice with longer tails and shorter hindfoot digits showed higher mean performance in an assay of arboreal locomotion. We further found that while some variance in climbing tendency may be due to variation in performance among mice, some hybrid mice engaged in extensive climbing behavior despite poor performance. In Chapter Three, we identified a large chromosomal inversion that was strongly associated with differences in both tail length and pigmentation, specifically hue. Using wild mice sampled across an environmental gradient between the focal western forest and prairie populations, we tested for associations among genotype, phenotype, and environment. We found that the pattern of allele frequency variation at the inversion differed from the genome-wide ancestry pattern, and the inversion showed elevated genetic divergence compared with the rest of the genome, suggesting that natural selection may be acting on the inversion. However, while genotype at the inversion was associated with tail length and hue in the laboratory and in the wild, clines for tail length and hue were coincident with both soil hue and genome-wide ancestry fraction, but not with the inversion. Thus, tail length and hue alone cannot explain the distribution of the inversion across the landscape. Together, we determined that many factors likely affect the evolution of the multiple forest phenotypes: phenotypic traits differ consistently by habitat in both eastern and western deer mouse lineages and for the most part, these traits are genetically unlinked and polygenic. However, aspects of skeletal morphology, such as tail length and hindfoot size, may be significant for performance differences among mice. In addition, a large inversion linked to tail length and pigmentation may contribute to local adaptation in western forest and prairie mice. Thus, this dissertation highlights the importance of understanding the patterns of genetic and functional linkage among traits to the study of local adaptation.